64 



SCIENCE. 



[Vol. XVIII. No 443 



poles in succession for exactly the same length of time (about 

 six seconds) ; by this means the variations in the strength of 

 the current, the experiment lasting some hours, affected each 

 metal alike. 



The exposed surface of the various metals used as nega- 

 tive poles was kept uniform by taking them in the form of 

 wires that had all been drawn through the same standard 

 hole in the drawplate, and cutting them by gauge to a uni- 

 form length; the actual size used was 0.8 of a millimeter in 

 diameter and twenty millimetres long. 



The comparison metal, gold, had to be used in each ex- 

 periment; the apparaitus thus enabled me to compare three 

 different metals each time. The length of time that the cur- 

 rent was kept on the revolving commutator in each experi- 

 ment was eiglit hours, making two hours of electrification 

 for each of the four negative electrodes; the pressure was 

 such as to give a dark space of six millimetres. 



The fusible metals, tin, cadmium, and lead, when put into 

 the apparatus in the form of wires, very quickly melted. To 

 avoid this difficulty a special form of pole was devised. Some 

 small circular porcelain basins were made, nine millimetres 

 in diameter; through a small hole in the bottom a short 

 length of iron wire, 0.8 of a millimetre in diameter, was passed, 

 projecting downwards about five millimetres; the basin was 

 then filled to the brim with the metal to be tested, and was 

 fitted into the apparatus exactly in the same way as the wires. 

 The internal diameter of the basin at the brim was seven 

 millimetres, and the negative metal filed iiat was thus formed 

 of a circular disk seven millimetres in diameter. The stan- 

 dard gold pole being treated in the same way, the numbers 

 obtained for the fusible metals can be compared with gold, 

 and take their place in the table. 



The following table of the comparative volatilities was in 

 this way obtained, taking gold as 100: — 



Palladium 108.00 1 Platinum 44.00 



Gold 100.00 1 Copper 40.24 



SUTer 82.68 I Cadmlam 31.99 



Lead 75.04 I Nickel 10.99 



Tin 56.96 I Iridium 10.49 



Braas 51.58 I Iron 5.50 



In this experiment equal surfaces of each metal were ex- 

 posed to the current. By dividing the numbers so obtained 

 by the specific gravity of the metal, the following order is 

 found : — 



Palladium 9.00 I Copper 8.58 



bilver 7,88 Platinum 8.02 



Tin 7.76 I Nickel I.a9 



lead 6.61 1 Iron 0.71 



»old 5.18 Iridium 0.47 



Cadmium 3.72 | 



Aluminum and magnesium appear to be practically non- 

 volatile under these circumstances. 



The order of metals in the table shows at once that the 

 electrical volatility in the solid state does not correspond with 

 the order of melting-points, of atomic weights, or of any 

 other well-known constant. The experiment with some of 

 the typical metals was repeated, and the numbers obtained 

 did not vary materially from those given above, showing 

 that the order is not likely to be far wrong. 



It is seen in the above table that the electrical volatility of 

 silver is high, while that of cadmium is low. In the two 

 earlier experiments, where cadmium and silver were taken, 

 the cadmium negative electrode in thirty minutes lost 7.53 

 grains, whist the silver negative electrode in 1^ hours only 

 lost 0.19 of a grain. This apparent discrepancy is easily ex- 

 plained by the fact (already noted in the case of cadmium) 

 that the maximum evaporation effect, due to electrical dis- 

 turbance, takes place when the metal is at or near the point 

 of liquefaction. If it were possible to form a negative pole 



in vacuo of molten silver, then the quantity volatilized in a 

 given time would be probably more than that of cadmium. 



Gold having proved to be readily volatile under the elec- 

 tric current, an experiment was tried with a view to produc- 

 ing a larger quantity of the volatilized metal. A tube was 

 made having at one end a negative pole composed of a 

 weighed brush of fine wires of pure gold, and an aluminum 

 pole at the other end. 



The tube was exhausted and the current from the induc- 

 tion coil put on, making the gold brush negative. The re- 

 sistance of the tube was found to increase considerably as the 

 walls became coated with metal, so much so that, to enable 

 the current to pass through, air had to be let in after a while, 

 depressing the gauge one-half of a millimetre. 



The weight of the brush before experiment was 35.494 

 grains. The induction current was kept on the tube for 

 14|^ hours; at the end of this time the tube was opened and 

 the brush removed. It now weighed 32.5613 grains, show- 

 ing a loss of 2.9327 grains. When heated below redness the 

 deposited film of gold was easily removed from the walls of 

 the tube in the form of very brilliant foil. 



After having been subjected to electrical volatilization, the 

 appearance of the residual piece of gold under the micro- 

 scope, using a quarter-inch object-glass, was very like that 

 of electrolytically deposited metal, pitted all over with mi- 

 nute hollows. 



This experiment on the volatilization of gold having pro- 

 duced good coherent films of that metal, a similar experi- 

 ment was tried, using a brush of platinum as a negative 

 electrode. On referring to the table it will be seen that the 

 electric volatility of platinum is much lower than that of 

 gold, but it was thought that by taking longer time a suffi- 

 cient quantity might be volatilized to enable it to be removed 

 from the tube. 



The vacuum tube was exhausted to such a point as to give 

 a dark space of six millimetres, and it was found, as in the 

 case of gold, that as a coating of metal was deposited upon 

 the glass the resistance rapidly increased, but in a much 

 more marked degree, the residual gas in the tube apparently 

 becoming absorbed as the deposition proceeded. It was 

 necessary to let a little air into the tube about every thirty 

 minutes, to reduce the vacuum. This appears to show that 

 the platinum was being deposited in a porous spongy form, 

 with great power of occluding the residual gas. 



Heating the tube when it had become this way non-con- 

 ducting liberated sufficient gas to depress the gauge of the 

 pump one millimetre, and to reduce the vacuum so as to 

 give a dark space of about three millimetres. This gas was 

 not re-absorbed on cooling, but on passing the current for ten 

 minutes the tube again refused to conduct, owing to absorp- 

 tion. The tube was again heated, with another liberation 

 of gas, but much less than before, and this time the whole 

 was re-absorbed on cooling. 



The current was kept on this tube for twenty-five hours; 

 it was then opened, but I could not remove the deposited 

 metal except m small pieces, as it was brittle and porous. 

 Weighing the brush that had formed, the negative pole gave 

 the following results: — 



Grains, 



Weight of platinum before experiment : 10.1940 



" " after experiment 8.1570 



Loss of volatilization In 25 hours 8 0370 



Another experiment was made similar to that with gold 

 and platinum, but using silver as the negative pole, the pure 

 metal being formed into a brush of fine wires. Less gas was 

 occluded during the progress of this experiment than in the 



